Beyond small molecules: advancing MYC-targeted cancer therapies through protein engineering

ABSTRACT

Protein engineering has emerged as a powerful approach toward the development of novel therapeutics targeting the MYC/MAX/E-box network, an active driver of >70% of cancers. The MYC/MAX heterodimer regulates numerous genes in our cells by binding the Enhancer box (E-box) DNA site and activating the transcription of downstream genes. Traditional small molecules that inhibit MYC face significant limitations that include toxic effects, drug delivery challenges, and resistance. Recent advances in protein engineering offer promising alternatives by creating protein-based drugs that directly disrupt the MYC/MAX dimerization interface and/or MYC/MAX’s binding to specific DNA targets. Designed DNA binding proteins like Omomyc, DuoMyc, ME47, MEF, and Mad inhibit MYC activity through specific dimerization, sequestration, and DNA-binding mechanisms. Compared to small molecules, these engineered proteins can offer superior specificity and efficacy and provide a potential pathway for overcoming the limitations of traditional cancer therapies. The success of these protein therapeutics highlights the importance of protein engineering in developing cancer treatments.

Transcription factors (TF) are DNA-binding proteins that regulate gene expression. They generally comprise two main domains: a transcriptional activation domain and a DNA-binding domain (DBD). The activation domain recruits transcription elements that enhance or repress gene transcription. The DBD specifically binds to DNA sites within DNA promoters. DNA promoters are blocks of 100–1000 base pairs (bp) positioned upstream or directly downstream from the gene. When bound to promoters, TFs regulate transcription of a specific gene [1–3].

Due to the importance of TFs in modulating the cellular proteome, TFs are in turn regulated on multiple levels: the translational level and the post-translational level. The translational level refers to the extent of cellular expression of a TF, while the post-translational level refers to the post-translational modifications that modulate the activities of TFs [1]. We can exploit the different levels of regulation to design drugs that specifically target – and then control – an aberrant TF’s activity [4,5].

Challenges with targeting transcription factors

Aberrant TF activity is often linked to various diseases. Yet TFs are often labeled “undruggable” because they lack the well-defined active or allosteric sites typically targeted in small-molecule drug design. These sites, often pockets or clefts on proteins, allow small molecules to bind and modulate protein function. DBDs are particularly difficult to target with small-molecule drugs. DBDs are rich in positively charged basic residues that would not typically interact with most small-molecule drugs, which are usually hydrophobic. Transcriptional activation domains are also challenging to target using small molecules. Activation domains are involved in protein-protein interactions and generally adopt a “flat” structure. In other words, activation domains lack the pockets needed for effective small-molecule drug design. Moreover, activation domains share a scaffold with other activation domains in the same class, making the targeting of individual, specific activation domains difficult [6]. Therefore, targeting TFs with small molecules has been especially difficult [7–9].

Nevertheless, scientists have been creative in approaching this issue. In this review, we will discuss recent therapeutic advances targeting MYC, a basic/helix-loop-helix/leucine zipper (bHLHZ) transcription factor. We will focus on recent protein-based drugs that target the MYC dimerization interface and its binding to DNA. We will also discuss the protein engineering approaches that were used to design these protein therapeutics.

The structure of bHLH modules

The basic/helix-loop-helix (bHLH) modules are a class of TFs found in eukaryotic organisms (reviewed in ref [10]). The bHLH domain comprises two main regions, a basic region and a helix-loop-helix region (HLH, Figure 1). The basic region contains a stretch of approximately 15 amino acids that is rich in basic residues. The basic region is typically disordered or only partially helical and can only become a fully helical extension of the HLH upon binding to DNA (nonspecific genomic DNA or its specific DNA target) or another polyanionic molecule [13–17]. Most bHLH motifs specifically bind a 6 bp DNA target site called the enhancer box (E-box), which comprises the sequence 5’-CANNTG, where N indicates any nucleotide. When bound to the E-box DNA response element, the basic region lies in the major groove of DNA and specifically interacts with the nucleotide bases and backbone phosphodiester groups of the E-box DNA site (Figure 2a) [18].

Figure 1.

Example of bHLH in complex with DNA. TCF4 homodimer bound to the E-box DNA (PDB: 6OD3) [11]. The four-helix bundle (two monomers shaded light and dark green) and basic regions (light and dark purple) bind to DNA (brown). One of the monomers is denoted with helices 1 and 2 while the other monomer is denoted with helices 1’ and 2’ [11]. Molecule viewed using ChimeraX v. 1.8 [12].

Figure 2.

Transcription factor 4 bound to E-box. TCF4 homodimer is an example of a bHLH protein that is essential for brain development. TCF4 bHLH monomers are shaded light and dark green (HLH) and light and dark purple (basic region) (PDB: 6OD4) [11]. The basic region lies in the major groove and makes specific contacts with the DNA. Panel a shows hydrogen bonds between the protein and the DNA major groove (dotted black line). The HLH is a four-helix bundle comprising two monomers making noncovalent interactions with each other. One such interaction is represented in panel b as a hydrogen bond (dotted black line). The N-terminus of helix 2 of the HLH makes stabilizing contacts with the phosphodiester backbone of the DNA, represented as hydrogen bonds in panel c (dotted black line). Molecules viewed using ChimeraX v. 1.8 [12].

While the basic region is responsible for DNA binding, the HLH facilitates protein dimerization. The HLH comprises two amphipathic helices separated by a loop. Helix 1 extends from the basic region and contains residues capable of forming specific protein-protein interactions with helix 1’, which is helix 1 within the HLH of the protein partner (Figures 1 and 2(b)). Helix 2 further stabilizes dimerization by forming a coiled coil with helix 2’ through hydrophobic packing. The N-terminal base of helices 2 and 2’ lies close to the DNA and can interact with the DNA phosphodiester backbone (Figure 2(c)). This packing of the four helices in the bHLH dimer allows formation of a four-helix bundle (Figure 1), which are four antiparallel α-helices that are held together mainly via noncovalent hydrophobic and van der Waals interactions among amino acid side chains [15,19]. A subset of the bHLH superfamily utilizes a leucine zipper (LZ) extending from helix 2 to create the basic/helix-loop-helix/zipper family (bHLHZ). The LZ reinforces protein dimerization partner specificity, allows for more selective homo- or heterodimerization as bHLHZs bind their DNA targets, and provides a larger interface for protein-protein interaction that can further stabilize the protein dimer (Figure 3) [20].

Figure 3.

Example bHLHZ in complex with DNA. MYC/MAX heterodimer bound to the E-box (tan) (PDB: 1NKP) [21]. MYC is shaded in dark colors, MAX in light colors. The basic regions are blue, HLH purple, and leucine zipper orange. Molecules viewed using ChimeraX v. 1.8 [12].

The function of MYC, a bHLHZ TF

The most well studied bHLHZ protein that regulates cell growth and proliferation is MYC, mainly due to MYC’s oncogenic properties and it being one of the first identified bHLHZ proteins [22,23]. MYC only dimerizes with MAX, another bHLHZ protein that serves as a master regulator of the bHLHZ subfamily. MYC must partner with MAX to bind the E-box DNA response element and function as a transcriptional regulator [24–27]. The MYC/MAX/E-box network is involved in the regulation of around 15% of normal gene activity [18,28–31]. Upon binding the E-box, the MYC/MAX heterodimer recruits various proteins, including coactivator complexes, to activate the expression of downstream proteins [32,33]. However, MYC has also exhibited some repressive abilities, so it is not purely a transcriptional activator [32,34–37].

MYC regulates the cell cycle through binding and regulating partner proteins. For instance, Miz-1 interaction with MYC/MAX plays a vital role in regulating downstream genes involved in cell cycle and apoptosis [38–40]. While P53 and MYC were termed the yin and yang of cell cycle regulation, as the balance of these two proteins play a vital role in cell cycle arrest and differentiation [41].

MYC also interacts with the E2F family to regulate cell cycle progression and apoptosis. Interaction between MYC and E2F2 or E2F3 induces cellular progression into the S phase, in which the cell replicates its genomic material, while interaction with E2F1 induces cellular apoptosis [42]. MYC regulates cellular levels of lactate dehydrogenase A (LDH-A), thereby mediating cellular differentiation and progression to the S phase [43–45]. MYC also regulates the levels of EIF5A, a protein elongation factor involved in the translation of proteins containing consecutive prolines: EIF5A is increasingly identified as playing a key role in regulating tumor cell growth (reviewed in refs [46–48]). Lastly, RNA Polymerase III transcribes the S55 ribosomal subunit and other non-coding RNAs, and the levels of these RNA transcripts are associated with cellular growth. MYC binds TFIIIB, an RNA pol III specific transcription factor, directly activating RNA Pol III transcription [49,50]. The MYC activities described above are just a subset of MYC’s various functions and the many ways MYC regulates the cell cycle.

In turn, MYC’s activity is regulated by many other proteins, most notably the MAX dimerization proteins (MXD). The MXD proteins dimerize with MAX and bind to the E-box, thereby impeding the formation of the MYC/MAX heterodimer, which is crucial for MYC’s activity as a DNA binder. Additionally, MXD has a mSin3 interacting domain that recruits mSin3, a transcriptional repressor that assists MXD’s repression of the regulated genes [30,51]. MXD also binds to UBF (upstream binding factor), a ribosomal DNA regulatory factor. MYC’s binding to UBF induces the transcription of ribosomal RNA (rRNA). By binding UBF, MXD sequesters UBF, downregulating rRNA transcription [52,53].

Due to the importance of the MYC/MAX network in regulating cell proliferation, it is of no surprise that misregulation in this network is associated with approximately 70% of human cancers [54–56]. Thus, the MYC/MAX transcription factor proteins have been a desirable target for anti-cancer drugs. Since MAX is a master regulator, drug-design efforts were typically targeted toward MYC. However, the druggability of MYC, MAX, and the MYC/MAX dimer poses many challenges, as mentioned above, namely, the lack of “pockets” in the bHLHZ domain that can be targeted by small molecules, the high charged-residue content, and the disordered nature of the basic region that makes drug design challenging [57–59]. Nevertheless, researchers have attempted to create small molecules and protein drugs that work to disrupt the MYC/MAX/E-box network, as discussed below.

Small molecules that indirectly target the MYC/MAX network: bromodomain inhibition

Researchers have pursued the development of small-molecule drugs that disrupt the MYC/MAX network through indirect mechanisms, such as targeting the extraterminal (BET) bromodomain family of proteins [60]. BET proteins – such as BRD2, BRD3, and BRD4–utilize their acetylated lysine recognition pocket to bind acetylated histones [61,62]. The binding of BET proteins to acetylated histones alters chromatin structure and enhances chromatin accessibility to transcriptional machinery [63]. BET proteins then recruit the positive transcription elongation factor complex b (P-TEFb), which facilitates the transcription of genes like MYC (Figure 4(a)) [66–69]. Therefore, inhibiting BET bromodomains can indirectly modulate MYC/MAX activity by disrupting MYC expression.

Figure 4.

Overview of bromodomains and bromodomain inhibitors. (a) Bromodomains recognize and bind to acetylated histones. This results in the unwinding of chromatin structure that increases chromatin accessibility to transcriptional machinery, which then facilitates the transcription of genes such as MYC. Bromodomain inhibitors, including JQ1 and OTX015, compete with acetylated histones for bromodomain binding sites. These inhibitors effectively inhibit bromodomain activity and subsequently suppress MYC expression. (Adapted from ref [64]). (b) Structure of JQ1, one of the earliest and most widely studied bromodomain inhibitors [65]. (c) Structure of OTX015 (produced by Merck), another potent bromodomain inhibitor. OTX015 shares structural elements with JQ1, thereby explaining the similarity in toxicity profiles and pharmacokinetic properties.

Targeting MYC via bromodomain inhibition presents two significant challenges. First, bromodomains are involved in regulating cellular pathways beyond MYC, such as NF-κB and NOTCH1 [70–73]. The NF-κB pathway drives inflammatory and immune responses, while the NOTCH1 signaling pathway drives the differentiation, development, and function of hematopoietic stem cells [72,74–76]. Inhibiting these pathways through bromodomain inhibitors can lead to drug toxicity, posing a risk to patients.

Second, a significant obstacle is the potential for resistant cancer cells to evolve by upregulating compensatory mechanisms that prevent MYC levels from decreasing despite the administered drug [77,78]. One critical pathway that cancer cells upregulate is the WNT/β-catenin signaling pathway. WNT, an extracellular protein, binds to its receptors, Frizzled and LRP5/6, on the cell membrane. The binding of WNT to Frizzled/LRP5/6 inhibits the proteolysis of β-catenin in the cytoplasm, allowing β-catenin to accumulate and translocate to the nucleus, where it activates transcriptional pathways associated with MYC [72,79–82]. The persistence of the WNT/β-catenin pathway and other pathways that regulate MYC emphasizes the necessity of combining bromodomain inhibitors with other agents to target these compensatory mechanisms and prevent the emergence of resistant cancer strains [78].

JQ1 (Figure 4(b)), a pioneering BET bromodomain inhibitor, has shown considerable promise for suppressing MYC-driven cancers in preclinical studies by targeting the acetylated lysine recognition pocket of BET proteins [65]. Recent in vivo studies in both tumor xenografts extracted from patients and mouse models have demonstrated that JQ1 can suppress the growth of MYC-driven cancers, such as endometrial cancer, lymphoma, and medulloblastoma [83–85]. These findings build on earlier research showing that JQ1 specifically suppresses MYC expression, leading to cell-cycle arrest and cellular senescence in multiple myeloma cancer cells, where MYC is the primary driver gene [69].

Despite these promising results, JQ1 shows limited efficacy in clinical trials. Patients often relapse due to the activation of compensatory mechanisms that sustain MYC levels [86–90]. Additionally, JQ1 exhibits toxic effects when administered at high doses (>30 mg/kg body weight, twice daily), including severe weight loss (over 10% overall body weight), decrease in hematological profiles (lymphocyte, monocyte, and total white blood cell counts), and immunological deficiencies involving low lymphocyte count [91]. Furthermore, JQ1 has a poor pharmacokinetic profile, characterized by high hydrophobicity (LogP 4.9) and low bioavailability (half-life of only 54 minutes) [91–93].

These limitations are not confined to JQ1 alone: other bromodomain inhibitors, such as OTX015 (produced by Merck, Figure 4(c)), have demonstrated similar disadvantages. Patients treated with OTX015 experienced adverse effects like anemia, fatigue, headaches, back pain, and neutropenia [72,93–96]. Moreover, like JQ1, OTX015-treated patients developed resistance to the drug [95]. These drawbacks severely restrict the clinical utility of bromodomain inhibitors and have prompted researchers to explore alternative strategies for targeting MYC-driven cancers.

Small molecules that directly target MYC and MAX

In recent years, researchers focused on developing small-molecule drugs to directly target MYC with the aim of reducing side effects associated with indirect MYC inhibition. These drugs specifically target the dimerization interface of MYC and MAX, as forming the MYC/MAX heterodimer is essential for MYC’s oncogenic function. For example, Castell et al. developed MYCMI-6, a molecule designed to bind MYC and disrupt its heterodimerization with MAX, thereby preventing MYC from binding to DNA [97,98]. Similarly, Struntz et al. created KI-MS2-008, a MYC/MAX inhibitor that targets MAX by stabilizing the MAX homodimer. This stabilization hinders MYC/MAX interaction and heterodimerization, effectively interfering with MYC’s oncogenic activity [99]. Other discovered molecules inhibit MYC’s activity in a manner similar to MYCMI-6 and KI-MSI2-008 [100,101].

Both MYCMI-6 and KI-MS2-008 have shown potential as therapeutic agents, demonstrating efficacy against MYC-dependent tumor lines. MYCMI-6 exhibits growth inhibition concentration (IC₅₀) values ranging from 2.5 to 6 µM in MYCN-amplified neuroblastoma cell lines [97]. KI-MS2-008 inhibits the growth of the P493–6 B-cell line that has been engineered with exogenous c-Myc, with IC₅₀ value 2.2 µM [99].

However, the efficacy of these molecules remains limited, as their pharmacodynamic values are still in the µM range, which is not ideal for drug development. The challenge also lies in the large dimerization interfaces of bHLHZ domains that make the bHLHZ difficult to target effectively with small molecules – hence, the inevitability of needing to use µM range dosages [102]. Additionally, the delivery of these small-molecule drugs to the nucleus – where bHLHZ proteins are active – poses further challenges, as ensuring a sufficiently high drug concentration in the nucleus is crucial for efficacy. Moreover, bHLHZ and related transcription factors are ubiquitous in the cell nucleus, complicating the selective targeting of MYC/MAX or MAX/MAX interfaces against other α-helical protein dimerization interfaces [103,104].

Given these limitations, it is essential to explore alternative strategies for targeting MYC. One promising avenue is the design of protein drugs that mimics the structure of bHLHZ proteins but possess altered dimerization domains. These engineered proteins can specifically sequester MYC’s activity by interacting as dimerization partners with bHLHZ proteins and by binding to the E-box site, thereby blocking MYC’s access to the E-box. These protein drugs do not contain any of the activation domains necessary to activate the MYC-associated pathway. Thus, these protein drugs are incapable of activating any pathway upon binding the E-box response element. Such strategies can offer a more effective approach to inhibiting MYC-driven tumorigenesis.

Omomyc: a MYC-based therapeutic

The Soucek group successfully engineered Omomyc – a small protein therapeutic designed to inhibit MYC activity [105–107]. Omomyc is the most successful developed protein therapeutic, having progressed to the clinical stage of testing [108], while other discussed protein-based therapeutics have not gone past the cellular and/or animal studies stage. Omomyc was specifically designed to heterodimerize with MYC. This was achieved by introducing four strategic mutations in the leucine zipper of MYC’s 91 amino-acid bHLHZ domain, as the LZ is the element that largely controls protein-partnering specificity [15,109,110]. The Soucek group rationalized that these mutations would now allow MYC to homodimerize, as MYC does not homodimerize due to electrostatic repulsion in the LZ region (Figure 5(a)) [105]. This electrostatic repulsion originates from two arginine residues at positions 70 and 71 and two glutamate residues at positions 57 and 64. In contrast, the MAX protein can homodimerize, because it has a glutamine and asparagine at positions 70 and 71, respectively, which form a stable tetrad that enhances the stability of the MAX homodimer’s four-helical bundle in the HLH (Figure 5(b)) [105,111]. To replicate the MAX bHLHZ’s stable tetrad structure in the MYC bHLHZ, the Soucek group introduced two mutations, R70Q and R71N, into MYC’s LZ. Additionally, the Soucek group made two other mutations, E57T and E64I, to improve shape complementarity between the dimerization interfaces of the LZ α-helices [105].

Figure 5.

The rationale behind the design of Omomyc. (a) the LZ structure within MYC’s bHLHZ (PDB: 1NKP) [21] shows that residues R70 and R71 (green), along with E57 and E64 (red and yellow, respectively), are responsible for the electrostatic repulsion that prevents MYC from homodimerizing [20,105]. (b) in the MAX homodimer crystal structure (PDB: 1AN2) [111], residues Q91 and N92 from each monomer interact to form a stable tetrad that reinforces homodimerization. The black dashed lines represent the interactions between the protein residues [105,111]. (c) in the MYC/MAX heterodimer crystal structure (PDB: 1NKP) [21], Q70 and N71 from the MAX LZ interact with R70 and R71 from the MYC LZ and further reinforce heterodimerization [21,112]. The black dashed lines represent interactions between protein residues. Molecules were viewed using ChimeraX v. 1.8 [12].

Surprisingly, although the four mutations were introduced into Omomyc to reduce electrostatic repulsion and facilitate heterodimerization with MYC, these mutations do not appear to facilitate the Omomyc/MYC heterodimer, as these heterodimers do not bind DNA with high affinity [105,113–115]. However, Omomyc demonstrates a stronger propensity to homodimerize and to dimerize with MAX [107,113]. In the MYC/MAX heterodimer, Gln70 and Asn71 of the MAX bHLHZ domain form strong hydrogen bonds with the positively charged Arg70 and Arg71 of the MYC bHLHZ domain, thereby facilitating MYC/MAX dimerization (Figure 5(c)) [21,112]. In Omomyc, these arginine residues are mutated to resemble those in MAX, allowing potential tetrad interactions between Omomyc and MAX, like those in MAX homodimers [105]. This also explains Omomyc’s capability to homodimerize, as this favorable tetrad structure would be present between two Omomyc monomers. In hindsight, Omomyc’s inability to dimerize with MYC is understandable, as MYC and MAX have evolved to function specifically as a heterodimer. Consequently, any protein scaffold based on MYC is more likely to interact primarily with MAX.

Given Omomyc’s protein-partnering preferences, Omomyc is believed to inhibit MYC activity by binding the E-box as a homodimer or as a heterodimer with MAX, thereby competitively inhibiting the MYC/MAX heterodimer from binding the E-box target site (Figure 6) [107,113]. Because both Omomyc and MAX lack a transcriptional activation domain [116], the Omomyc homodimer and the Omomyc/MAX heterodimer do not activate MYC-associated pathways upon binding DNA. This dual mechanism of action involves competing with the MYC/MAX heterodimer for access to the E-box while simultaneously disrupting MYC/MAX heterodimerization by sequestering free MAX available for dimerization with MYC [107,113]. Since Omomyc has a stronger tendency to dimerize with MAX rather than homodimerize with itself, the Omomyc/MAX heterodimer binding to the E-box is the primary complex responsible for Omomyc’s inhibitory effect on MYC [114,115].

Figure 6.

Omomyc’s mechanism of action. (a) Omomyc homodimer binds to the E-box, thereby preventing the MYC/MAX heterodimer from binding to the E-box and effectively blocking MYC activity [105,107,113]. (b) Omomyc heterodimerizes with MAX and prevents the formation of the MYC/MAX heterodimer, which is crucial for MYC’s activity [25]. MYC is only active as a transcriptional activator when partnered with MAX. The Omomyc/max heterodimer binds to the E-box, thereby inhibiting MYC activity [105,107,113].

DuoMYC

More recently, DuoMYC has emerged as a potential therapeutic for MYC-driven cancers [117]. DuoMYC is a synthetic miniprotein comprising two Omomyc bHLH monomers connected by a synthetic linker at the C-terminus of the HLH motif. Since the synthetic linker fuses the two Omomyc monomers into a single stable complex, without the use of a LZ, DuoMYC mimics the Omomyc homodimer [107,113]. Electrophoretic mobility shift assays (EMSA) with DuoMYC showed its capability to bind the E-box with affinities in the low 100’s nM, while in cellulo experiments displayed DuoMYC’s ability to counteract the gene mis-regulation caused by MYC overexpression. DuoMYC is proposed to compete with MYC/MAX binding to the E-box, thereby inducing its therapeutic action. Although early in its studies, DuoMYC exhibits promising results and could be a potential therapeutic or a scaffold for the design of future MYC-driven cancer therapeutics [117].

ME47 and MEF: max-based therapeutics

The Shin group engineered ME47 (formerly published as MAX-E47) by fusing the basic region from the MAX bHLHZ with the HLH subdomain of the E47 bHLH TF [118,119]. MAX belongs to the bHLHZ transcription factor family, while E47 is a member of the bHLH family; the bHLHZ and bHLH are different transcription factor families, despite having structural similarities [15,109,110].

E47’s ability to homodimerize stems from its structure (Figure 7). E47’s helix 1 within the HLH is one turn longer than helix 1 of MAX’s bHLHZ domain. The additional amino acids within the helix increase the number of contacts within the dimerization interface of E47 monomers [19]. Specifically, a salt bridge forms between His366, which is near the C-terminus of helix 1, and Glu390, which is at the C-terminus of helix 2 of the dimerization partner (Figure 7(a)) [19,118,119]. This characteristic is shared by E proteins, such as E47 and E12, which lack a LZ element – a feature that typically reinforces the dimerization function in other protein motifs [19,121].

Figure 7.

The ME47 bHLH crystal structure (PDB: 3U5V) [120]. ME47 comprises MAX bHLHZ’s basic region and the E47 bHLH’s HLH module [118–120]. (a) the extra turn in helix 1 within ME47’s HLH (or E47’s HLH) results in an additional interaction between H366 and E390 in helix 2’ that further stabilizes the four helical bundle [19,118,119]. (b) the glutamine triad (Q364, Q373, and Q381) forms a hydrogen bond network with G360 that further stabilizes E47’s four-helix bundle and enforces stability, thereby allowing ME47 (which contains the E47 HLH) to crystallize as a dimer without the need for a DNA ligand to induce bHLH dimerization [19,118,119]. Molecules were viewed using ChimeraX v. 1.8 [12].

ME47 is capable of dimerizing in the absence of DNA [120], a characteristic that distinguishes ME47 from the MYC/MAX heterodimer and the MAX homodimer [21,111]. This unique ability is likely due to the overall stability of the E47 HLH domain featuring a glutamine triad comprising Gln373, Gln364, and Gln381 that forms a hydrogen bond network with Gly360 (Figure 7(b)) [120]. This network further stabilizes the orientation of the helices within each monomer, reinforcing E47’s dimerization interface. As a result, two ME47 monomers can dimerize without the presence of DNA [19,118,119]. Crystallographic studies demonstrate that ME47 can crystallize as a stable dimer in the absence of a DNA ligand [120], whereas the MYC/MAX heterodimer and the MAX homodimer require the E-box DNA ligand to form stable crystallized structures [21,111].

ME47 only homodimerizes. ME47 does not dimerize with MYC and is unlikely to dimerize with endogenous E47 or any proteins with which E47 heterodimerizes, such as MyoD [118,122]. This is because ME47 is a chimera that contains modules from different protein families – the bHLH and the bHLHZ families – that exclusively partner with members of their own families [15,109,110]. Thus, ME47 inhibits MYC’s activity by competitively binding the E-box as a homodimer (Figure 8) [118,119].

Figure 8.

Repression of MYC activity by ME47. ME47 binds as a homodimer to the E-box, thereby competing with MYC/MAX binding to the same site and inhibiting MYC’s activity [118].

The next-generation derivative of ME47 is MEF [123], which is a bHLHZ protein that comprises ME47 (with Arg12 and Cys29 mutated to Ala for enhanced E-box binding affinity and protein stability) fused to FosW [124,125], a semi-rationally designed LZ modeled after the c-fos basic region-leucine zipper (bZIP) protein. Due to MEF’s dimerization interface, which possesses both the HLH of a bHLH protein and the LZ of a bZIP protein, neither MEF nor its derivatives are expected to partner with any endogenous proteins within the cell, as MEF would not be a member of any transcription factor family, including bHLH, bHLHZ, and bZIP.

MAX STR

The Moellering lab has created a synthetic transcriptional repressor (STR) derived from the MAX bHLHZ. This STR mimics the structure of the MAX bHLHZ domain; it comprises the basic region and minimal LZ chemically ligated together to mimic the overall bHLHZ tertiary structure. MAX STR was developed as a scaffold for future STR design. The STR was shown to homodimerize and proposed to block aberrant MYC activity by competitively binding to the E-box. EMSA experiments displayed MAX STR’s ability to maintain low nM affinity and specificity to the E-box [126].

MYC/MAX dimer-based therapeutics

The Pentelute group used flow-based protein synthesis to make dimeric transcription factor analogs. They stapled Omomyc, MYC, and MAX monomers as homo- or heterodimers by using chemical linkers. Through this attachment, the proteins would constantly exist as complexes that are capable of binding to DNA. These “dimeric” proteins showed DNA binding in EMSA experiments and significant solution stability compared to the monomers in circular dichroism experiments. In cellulo experiments showed the ability of the covalently linked proteins to interfere with MYC-driven gene expression [127,128].

Mad: a MXD1-based therapeutic

The Demma group engineered the protein Mad by utilizing the first 146 amino acids of MXD1, which include both the bHLH domain and the mSIN3a interaction domain (SID). The SID engages with mSIN3a, a transcriptional repressor that further inhibits MYC-driven gene transcription [129]. The Demma group introduced a serine-to-alanine mutation at position 145 to prevent phosphorylation, which is typically mediated by RSK and S6K1 kinases upon serum or insulin stimulation [129,130]. Phosphorylation at Ser145 accelerates the ubiquitination of MXD1, leading to MXD1’s degradation via the 26S proteasome pathway [130]. The degradation of MXD1 increases MYC activation, because MXD1 is no longer available to interact with MAX, thereby allowing the formation and activation of the MYC/MAX heterodimer [130]. Since Mad is designed to inhibit MYC activity, the serine-to-alanine mutation at position 145 eliminates the phosphorylation site, which should enhance the stability of Mad in the cell and potentially increase its efficacy as a therapeutic agent [129].

Mad’s mechanism of action closely follows that of MXD1. Mad heterodimerizes with MAX and binds to the E-box, effectively inhibiting MYC activity [129]. Mad also interacts with upstream binding factor (UBF), disrupting MYC’s ability to stimulate ribosomal gene transcription. Furthermore, the SID module within Mad engages with mSIN3a to further inhibit MYC-driven gene transcription. Therefore, Mad operates through three distinct mechanisms: 1) Mad heterodimerizes with MAX and binds to the E-box, thereby displacing the MYC/MAX heterodimer and suppressing MYC activity (Figure 9(a)); 2) Mad interacts with UBF and inhibits ribosomal gene transcription typically activated by MYC (Figure 9(b)); and 3) the SID within Mad recruits mSIN3a to repress MYC-driven gene expression (Figure 9(c)) [129].

Figure 9.

Repression of MYC activity by Mad. Mad contains the MXD1 bHLH domain and the mSin3a interaction domain (SID). Mad represses MYC activity through three mechanisms. (a) the Mad bHLH domain dimerizes with MAX and prevents the formation of the MYC/MAX heterodimer. The Mad/max heterodimer then binds to the E-box to effectively block MYC activity. (b) the SID interacts with mSin3a, thereby leading to the transcriptional repression of genes normally activated by MYC. (c) Mad interacts with upstream binding factor (UBF) to inhibit ribosomal gene transcription, which is also driven by MYC [129].

Omomyc, ME47, MEF, and Mad: pharmacodynamic and pharmacokinetic parameters

Omomyc, ME47, MEF, and Mad appear to exhibit superior efficacy as protein-based therapeutics compared to their small-molecule counterparts, particularly in targeting MYC-driven tumors. In vitro fluorescence polarization (FP) assays show that Omomyc demonstrates an affinity of 22 nM to the E-box, while in vitro EMSAs show that ME47 exhibits an even stronger affinity of 15.6 nM [113,118]. EMSA experiments also show that MEF has an E-box affinity that is half that of ME47, at 7.8 nM [123].

Although the dimerization affinity of Mad to MAX has not been explicitly quantified, FP DNA binding assays indicate that Mad DNA binding activity is similar to that of Omomyc [113]. This suggests that Mad’s affinity is likely acting within the same nM range as Omomyc [102]. Cell proliferation assays show Mad is 10 times more potent than Omomyc in inhibiting the growth of human colorectal carcinoma cells (HCT116) [129]. These results suggests that the SID module can be a powerful contributor toward inhibition of MYC-driven diseases [102]. These binding affinity values are three orders of magnitude more effective than those observed for direct protein inhibitors, whose affinities and inhibitory concentrations typically lie in the µM range [97,99]. These pharmacodynamic properties underscore the substantial potential of these proteins as therapeutic candidates against MYC-driven malignancies.

Beyond efficacy, the pharmacokinetic profiles, particularly the half-life and cellular stability of Omomyc, ME47, MEF, and Mad, are critical. While the half-lives of ME47, MEF and Mad have yet to be determined experimentally, insights can be inferred from the stability of the endogenous proteins they are modeled after. The half-life of these proteins can be related to the specific bHLH/bHLHZ domain that they contain. For instance, native MYC and MXD1 proteins are tightly regulated, with short half-lives of approximately 20–30 minutes, reflecting their roles in critical cellular processes [18,131,132]. Given the structural similarity between Mad and MXD1, it is reasonable to expect that Mad would exhibit a comparable half-life. Notably, Mad incorporates the S145A mutation, designed to reduce proteolysis, which may extend Mad’s half-life [129].

Similarly, Omomyc, which also shares structural features with MYC, has a reported half-life of 20–30 minutes [107,113]. However, advanced label-free mass spectrometry studies by the Soucek group have shown that Omomyc remains stable when administered intravenously in mouse models for up to 72 hours [133]. OMO-103, a proprietary formulation of Omomyc, demonstrated a terminal half-life of approximately 40 hours in a successful Phase I clinical trial. This was determined through pharmacokinetic (PK) studies conducted during dose-escalation, which involved analyzing serum samples and tissue biopsies collected from patients [108].

In contrast, ME47 and MEF are anticipated to exhibit greater stability than their counterparts, largely due to the nature of the endogenous proteins MAX and E47, upon which ME47 and MEF are based [118,119]. MAX, a master regulator within the cell, is neither abundant nor highly regulated, with a half-life of about 24 hours [18,131,132]. Similarly, E47, another master regulator, has a half-life of 6–8 hours [134]. Consequently, one may infer that ME47 and MEF may possess longer in vivo lifespans than Omomyc or Mad: this property could enhance their potential as long-lasting therapeutic agents [102].

Given these observations, selecting an appropriate bHLH or bHLHZ scaffold based on a robust protein is critical in designing therapeutic proteins. A robust scaffold not only enhances the stability and functionality of the resulting therapeutic protein but also simplifies the process of modification and optimization of a new therapeutic protein. This, in turn, improves the drug’s half-life, pharmacokinetic, and pharmacodynamic properties, ultimately ensuring it is both effective and safe for clinical use.

Limitations of protein therapeutics and potential modalities

There are several challenges to consider when implementing protein therapeutics – particularly those based on bHLHZ transcription factor scaffolds – as effective drugs. Protein therapeutics that mimic the long and extended structure of bHLHZ TFs are more susceptible to proteolysis compared to compact and globular proteins, as the loops of bHLHZ proteins are exposed to the extracellular environment [135]. Another major limitation of protein therapeutics is their poor oral bioavailability. This is due to enzymatic activity in the stomach that degrades protein therapeutics and the poor absorption of these proteins by the gastrointestinal tract, resulting in only 2% of the administered dose successfully bypassing the gastrointestinal barrier [136,137]. Consequently, protein therapeutics are often administered parenterally, a route associated with lower patient compliance compared to oral administration [137]. Even when administered parenterally, protein therapeutics often exhibit low distribution profiles due to limited cell and nuclear permeability [56,117]. These challenges emphasize the need for innovative delivery strategies and scaffold designs to enhance the stability, bioavailability, and overall efficacy of protein-based therapeutics.

To limit the effects of proteolysis, the peptide backbone can be modified by methylating the amino group of the backbone or introducing D-amino acids into the peptide [56]. Proteases degrade proteins by cleaving the peptide bonds holding the protein’s backbone intact, so introducing modifications that reduce the protease’s accessibility to the protein’s backbone – either by introducing steric hindrance or altering the stereochemical orientation of the backbone – can effectively limit protease activity on the protein therapeutic [138].

To improve the bioavailability of protein therapeutics, several strategies have been developed, including PEGylation and prodrug formation. PEGylation involves the attachment of polyethylene glycol (PEG) to the peptide therapeutic, thereby increasing the half-life of the protein therapeutic by creating steric hindrance against proteolytic enzymes. This modification helps protect the protein from enzymatic activity within the digestive system. Additionally, PEGylated proteins possess a larger surface area, which enhances their interaction with the intestinal wall, thereby increasing the likelihood of absorption [137,139,140]. Prodrug formation involves chemically modifying the original drug, often through esterification of hydroxyl, amino, or carboxylic acid groups within the molecule. This modification increases lipophilicity, thereby enhancing the drug’s permeability through the intestinal lining [137,141]. These approaches collectively aim to overcome the barriers to efficient delivery of protein therapeutics.

A potential solution that can mitigate both proteolysis and the poor delivery of protein therapeutics is the use of lipid nanoparticles (LNPs). LNPs act as delivery vehicles that transport molecules such as RNA and small-molecule drugs into cells. They achieve this through mechanisms such as clathrin- or caveolae-mediated endocytosis. If positively charged, LNPs can fuse with the negatively charged endosomal membrane, allowing their contents to access the cytoplasm [142].

Although LNPs have not traditionally been used to deliver protein therapeutics, LNPs offer potential for delivering either the protein itself or RNA segments encoding the protein, enabling its expression within cells [102]. However, certain challenges need to be addressed in the use of LNPs. These include LNPs lack of specificity in targeting cell types and their tendency to accumulate in the liver, thereby leading to toxicity. Despite limitations, LNPs provide a promising starting point for enhancing the drug delivery capabilities of protein therapeutics [143,144].

Conclusion

Targeting the MYC/MAX network through small molecules and protein-based therapeutics presents promising avenues for combating MYC-driven cancers. While small molecules like JQ1 have shown potential in preclinical studies, their limitations – such as toxicity, poor pharmacokinetics, and the emergence of compensatory mechanisms – emphasize the need for more effective strategies. Protein-based therapeutics – such as Omomyc, DuoMyc, ME47, MEF, and Mad – offer a more targeted approach, with significantly higher affinities to targets and the ability to disrupt MYC function using multiple mechanisms of action. These engineered proteins not only demonstrate superior efficacy in binding to the E-box and inhibiting MYC-driven transcription, but they also highlight the potential for overcoming the challenges posed by small-molecule inhibitors. As research progresses to minimize the limitations associated with protein therapeutics, these therapeutics could represent a new frontier in the treatment of MYC-driven malignancies and offer hope for more effective and less toxic cancer therapies.

Funding Statement

The work was supported by the NSERC.

Disclosure statement

No potential conflict of interest was reported by the author(s).